We capitalize on previously recorded kinematic and acoustic data for three languages (Georgian (GE), Spanish (SP) and Moroccan Arabic (MA)) that exhibit open articulatory transitions between the consonants in clusters in order to dissect the aero-acoustic parameters of the transitions in each language. These particular languages are of interest because they show similar patterns of interarticulatory timing in clusters, offering the unique opportunity to examine the acoustics of open transitions cross-linguistically. Our analysis centers on word initial clusters (/kl/ and /gl/), from which we extract relativized temporal values relevant to clusters and spectral parameters related to open articulatory transitions. We report baseline results using linear mixed effects models, then train a Random Forest model in a supervised learning environment on the significant variables. After training, test tokens are introduced in order to test whether the model can categorize the language based on the spectral and temporal parameters, and rank variables in terms of their feature importance. The results show that the model can categorize the data to the correct language with a 95.59% accuracy rate, where normalized zero-crossing (nzcr), modifications of the amplitude envelope (ΔE), and intensity ratio ranked highest in feature importance.